Predicting Business Cycle Turning Points with Neural Networks in an Information-Poor Economy
George Nasr, Ghassan Dibeh and Antoine Achkar
Summer Computer Simulation Conference 2007 (SCSC 2007)
San Diego, California (USA), July 15-18, 2007
Abstract
A feedforward neural network model is used to forecast turning points in the business cycle of postwar Lebanon. The NN has as inputs seven indicators of economic activity and as output the probability of a recession. The three-layered network is estimated using the back propagation algorithm. The NN is then used to forecast recursively a half-year ahead the probability of a recession in that period. The NN shows that two of the economic indicators can be used to construct a composite index of leading indicators that can be used to predict business cycles in the future.